Curious Speed + Time change on HP50G with Emu48 on Samsung

[Gil]

Try this small program on HP50G
with EMU48 on your phone :

« TIME
1. 100000.
START
NEXT
TIME
SWAP
HMS-
HMS->
3600 *
"Time in s"
->TAG »

Execution's time for these START/loops :
about 32 s with my Samsung Galaxy S6.

Now, I change the loop upper limit
from 100 000 to 1 million
(10 times more START/loops).
The execution's time jumps now to 1750 seconds instead of the expected 320 seconds. Why?

More curious.
Time shown on the calculator, after the 1 million START/loops, is equal to effective time + 89 minutes : the calculator shows a time about 1 hour 19 minutes ahead of effective time.

And even more unexpected:
After a while (of idleness) ,the calculator returns by itself to the correct time.

Any explanation and way to correct these strange behaviours?

Gil Campart

[Hlib]

Tested on go49gp emulator which works identical to real HP-50G with speed setting ×4 to make closer to your 32 s. It`s alright. I`m wondering: is this a problem in the EMU48 or in the Samsung Galaxy S6?

http://images.vfl.ru/ii/1576362127/f1258...912232.png

[Gil]

I'll test tomorrow evening with EMU48 on my son's Samsung Galaxy S10.

Very strange that the clock belonging to HP50G-EMU48 was temporarily changed (but of course not the clock of the phone itself), giving a calculation time of more than 6000 seconds, when the real time needed for those empty loops was about 9 minutes if I remember correctly (sure : more than 5 minutes, vs the 30-35 seconds expected).

I mentioned the "case" to Regis Cosnier and wait for his feedback.

Regards,
Gil

[Gil]

Precision or correction :
a) f(100 000 loops ) = 3.4 s;
b) f(1000 000 loops ) = 33 s;
c) f(10 000 000 loops) = 480 s
(expected about 330 s).

For c), the clock of HP50G with EMU48 on my Samsung Galaxy is systematically set ahead : last "run", about 1 h 38 m.

But as mentioned, the HP50G clock turns back by itself to the correct time of the phone.

Regards.

Gil